Site: NTT Optoelectronics Laboratories
Nippon Telegraph and Telephone Corporation
Tokai-Mura, Naka Gun
Ibaraki-ken, 319-11

Date Visited: November 15, 1994

Report Author: S. Forrest

ATTENDEES

JTEC:

D. Crawford
S. Forrest
R. Hickernell
F. Leonberger
C. Uyehara

HOSTS:

Kuniharu Kato

Associate Manager, Research Planning Division

M. Horiguchi

S. Mitachi

S. Sudo

T. Kitagawa

S. Imamura

K. Okamoto

Y. Ohmori

BACKGROUND

NTT's Optoelectronics (OE) Labs is one of several research laboratories that serve as the research arm of NTT. (General background information on NTT is included in the previous NTT site report.) The NTT research staff has remained relatively unchanged for the past ten years, while the development group has increased from less than 3,000 in 1986 to well over 5,000 today. Overall, NTT demonstrates a sizable commitment to both research and development, and it operates some of the premier industrial research labs in Japan.

The influence of NTT Labs on OE research and development in Japan is clearly evidenced by the ability of NTT engineers and scientists to often "set the agenda" for those in other Japanese companies. That is, problems first identified at NTT are often picked up as important issues and topics to be addressed at other Japanese OE companies at a later date. NTT's major optoelectronic research activities are centered at the Atsugi labs (see previous NTT site report by M. Warren), where both OEICs and other photonic devices are investigated. Activities at Ibaraki concentrate on photonic components and materials. There are roughly 250 researchers in these OE labs, with approximately equal numbers at the Atsugi and Ibaraki locations. The visit of the JTEC team covered several areas, including discussions of "hot" research topics, intellectual property issues, and technology transfer.

DISCUSSION

The JTEC panel was given presentations on several key technologies now under investigation at the labs. The technologies were planar lightwave circuits (PLCs) and other guided wave devices, optical amplifiers operating at both 1.3 microns and 1.55 microns wavelengths, connector technology, and polymer waveguide technology. In most cases, the activities could be classified as advanced development or engineering research. The exception was the connector technology, which was much closer to traditional product development work.

It is interesting to note that although there was considerable recognition by the NTT scientists of the need for cost reduction, there was no manufacturing research underway on photonic devices. That is, research directed primarily at improving manufacturing processes, including automated packaging, did not seem to be emphasized at the labs.

In the PLC research, the JTEC team was introduced to NTT's effort at fabricating PLCs on both Si and SiO ₂ substrates using a flame-hydrolysis waveguide deposition technique. Once the deposition of the doped silica has been accomplished, the films and substrate are heated in a furnace to achieve a dense, optical-quality waveguide. Using this process, 144 x 144 star couplers were fabricated with a high degree of uniformity from guide to guide. Optical coupling of the guides was accomplished by attaching fiber ribbons in groups of eight using Si v-groove assemblies. The JTEC team was told that the alignment of the v-groove assemblies is done using automated alignment tools.

A second guided wave device was a sophisticated thermo-optic matrix switch. This switch uses the temperature dependence of the refractive index of silica to switch 2 x 2 crossbars. The switch had a 2 ms switching time and a 15 dB on-off ratio employing a 0.5 W heater. Excess loss of 10 dB at 1.3 microns was claimed. The slow response of these large switching matrices was asserted not to be of consequence, since they were primarily needed for telephone line redundancy applications.

A final waveguide device was a dispersion equalizer consisting of 12 MZ (Mach-Zehnder) interferometers in a series. This device was designed to cancel fiber dispersion in 50 km of single-mode fiber. The PLC activity has 30 people in all, with approximately 10 of those concerned with packaging issues.

The most impressive research result was in the area of optical amplifiers. NTT claims to be the first lab to demonstrate a 1.3 microns fiber amplifier module. In fact, such a module used in transmitting a television signal was demonstrated to us during our visit. The core of the amplifier is a Pr-doped fiber pumped by a 1.017 microns solid state laser. The pump power was 340 mW. One particular problem with this technology is that a comparable semiconductor laser is still not available, although NTT is working on this device. The operating parameters of the amplifier were quite impressive. A noise figure of 3.4 dB is achieved that is superior to EDFAs, since the PDFA relies on a transition at 1.3 microns that is not at the ground state of Pr, which is in contrast to the process for the EDFA. Hence, re-excitation of the relaxed Pr atoms by the 1.3 microns light is not a problem contributing to noise. The amplifiers also have a gain of from 30 to 35 dB, with a gain efficiency of 0.4 dB/mW, and a saturated output power of +9 dBm.

The optical amplifier group was working at the time of the JTEC visit on Nd-doped fiber amps for use at 1.4 and 1.6 microns. These wavelengths may eventually be used for maintenance and supervisory functions in fiber-optic communication systems.

The most "engineering-oriented" activity we saw was in the fiber connector area. A full range of plastic, low cost single-mode connectors had been developed, including high precision panda fiber (for polarization sensitive applications) connectors. These friction assembled connectors have a polarization crosstalk of <-25 dB, and are very rapidly assembled using ferrules with key-ways to assure reproducible connections. Miniature connectors for use on crowded back-planes, with dimensions of only a few millimeters on a side have been developed. Perhaps the most noteworthy part of this work is that the development was carried well beyond the point where only the connector technology was perfected. Rather, a range of clever tools for inserting and demounting miniature connectors in the back planes, as well as connector and back plane cleaning tools for use in the field by technicians, were developed and demonstrated.

The final technology that NTT hosts showed the JTEC team was in the area of polymer waveguides. The developments at NTT in this emerging technology were concentrated in making very low-loss polymer waveguides on large substrates for a variety of fiber-optic applications. By replacing the C-H bond by its deuterium equivalent (i.e., C-D) in PMMA, a loss of 0.09 dB/cm was achieved at 1.3 microns, and 1.5 dB/cm at 1.55 microns. The 1.3 microns value meets the NTT-stated criterion for maximum allowable loss of 10 dB/m. On the down side of this result is a thermal stability of the material up to only 80 deg. C. Also, to achieve lower loss at 1.55 microns, deuterated polysiloxane was used. This had a loss of only 0.5 dB/cm at the longer wavelength, with an improved thermal stability. For example, no propagation loss increase was observed in the polysiloxane material after 100 h exposure to 120ÁC. The polymer activity is done by a staff of approximately 10 to 12 people.

Technology transfer of ideas generated at NTT laboratories to the public sector occurs via a three step process. In the first step, the technology is developed to a reasonable maturity at the labs. Next, this concept is transferred via technical papers and other documentation to NTT's R&D subsidiary; NTT Advanced Technology Corp. (NATC). Once the idea has been fully matured at NATC it is again transferred via papers and other documentation to one or more commercial manufacturers. Note that this transfer occurs without the transfer of personnel, although laboratory engineers will often interact with their counterparts at NATC to expedite the transfer process. Generally, only technologies not important to the NTT business sectors are transferred in this manner. The time scale for technology development within the NTT labs is 2 years or longer, depending on the complexity of the technology being considered. The technology is transferred from NTT to NATC after the development in NTT, and the time sale of the transfer is usually half a year to one year. Also, for this process to take place, NATC must pay a licensing fee to NTT, and the ultimate product manufacturer must in turn pay such a fee to NATC. [See NTT/Atsugi site report for further information on technology transfer.]

SUMMARY

It was clear that NTT labs continues to be the leading industrial laboratory in Japan. The quality and quantity of research projects, of which the JTEC team only saw a small fraction due to limited time, is extremely impressive. For comparison purposes, NTT can be considered the "Bell Labs" of Japan. Its activities appear to have considerable influence in setting the research agenda with regards to telecommunications systems elsewhere in the country. The sizable investment in research and development placed in these labs by NTT appears to be paying off considerable dividends in terms of providing the company with a strong, overall position in fiber optics long haul communications technology.